Localization of a Mobile Robot

OData support
Supervisor:
Kiss Domokos
Department of Automation and Applied Informatics

The Automated Driving is one of the most important direction of development nowadays in the automobile industry. To reach the goal of having driverless vehicles available, where there is no need for driver as a supervisor, the industry has to reach a lot of different milestones.

Four main topics have to be solved to reach the milestones:

• Where am I? The issue of localization.

• What is around me? Object detection

• What is happening next? Prediction

• How shall I react? Reaction, path planning

The main goal of my thesis was to investigate the first topic, which is the issue of the localization.

I had to select a proper robotic platform to reach the objective. The first basic requirement for the robotic platform is that the vehicle has to be a four wheeler. The second is regarding the remote controlling. The developer shall be able to remote control the vehicle to perform tests.

I implemented the initialization and handling of the sensors on the robotic platform. After that, I recorded the measurements data with different test setups. I processed the raw sensor data and I tried to compensate the typical errors for each of the sensors. During the correction I used different approaches and I compared them.

After processing the data, I implemented the wheel based odometry and inertial navigation and I compared the advantages and disadvantages of the different navigation models through the movement of the vehicle

At the end of my work with the help of the comparison I suggested a solution for the mobile robot localization problem based on different sensor fusion methods. I tested the compared the result of the fusion with the ground truth

Downloads

Please sign in to download the files of this thesis.